On 10/28, Chao Yu wrote:
> On 2022/10/28 11:22, Jaegeuk Kim wrote:
> > On 10/28, Chao Yu wrote:
> > > On 2022/10/25 14:50, Yangtao Li wrote:
> > > > Added a new sysfs node called gc_urgent_idle_remaining.
> > > > The user can set the trial count limit for GC urgent and
> > > > idle mode with this
Hello,
Thanks for the report, the bug has been fixed w/ below patch:
https://lore.kernel.org/linux-f2fs-devel/20221021023422.3101585-1-c...@kernel.org/T/#u
On 2022/10/28 11:34, syzbot wrote:
syzbot has bisected this issue to:
commit 4d67490498acb4ffcef5ba7bc44990d46e66a44c
Author: Fengnan
On 2022/10/28 11:22, Jaegeuk Kim wrote:
On 10/28, Chao Yu wrote:
On 2022/10/25 14:50, Yangtao Li wrote:
Added a new sysfs node called gc_urgent_idle_remaining.
The user can set the trial count limit for GC urgent and
idle mode with this value. If GC thread gets to the limit,
the mode will turn
On 2022/10/28 11:17, Jaegeuk Kim wrote:
On 10/28, Chao Yu wrote:
On 2022/10/28 10:33, Yangtao Li wrote:
It won't pollute global namespace since it's a static function just be used in
f2fs/super.c...
emm, I think it would be nice to see the f2fs_record_error_work symbol
in the stack, it can
syzbot has bisected this issue to:
commit 4d67490498acb4ffcef5ba7bc44990d46e66a44c
Author: Fengnan Chang
Date: Thu Aug 19 08:02:37 2021 +
f2fs: Don't create discard thread when device doesn't support realtime
discard
bisection log:
On 10/28, Chao Yu wrote:
> Jaegeuk,
>
> There is a generic fix [1], let's drop this one...
Ok, thanks.
>
> https://lore.kernel.org/lkml/20221025071549.1280528-1-yangyingli...@huawei.com/T/
>
> On 2022/10/20 16:45, Chao Yu wrote:
> > On 2022/10/18 15:32, Yang Yingliang wrote:
> > > Inject
On 10/28, Chao Yu wrote:
> On 2022/10/25 14:50, Yangtao Li wrote:
> > Added a new sysfs node called gc_urgent_idle_remaining.
> > The user can set the trial count limit for GC urgent and
> > idle mode with this value. If GC thread gets to the limit,
> > the mode will turn back to GC normal mode
On 10/28, Chao Yu wrote:
> On 2022/10/28 10:33, Yangtao Li wrote:
> > > It won't pollute global namespace since it's a static function just be
> > > used in f2fs/super.c...
> >
> > emm, I think it would be nice to see the f2fs_record_error_work symbol
> > in the stack, it can be explicitly a
On 2022/10/28 10:33, Yangtao Li wrote:
It won't pollute global namespace since it's a static function just be used in
f2fs/super.c...
emm, I think it would be nice to see the f2fs_record_error_work symbol
in the stack, it can be explicitly a function of f2fs. personal opinion for
reference
> It won't pollute global namespace since it's a static function just be used
> in f2fs/super.c...
emm, I think it would be nice to see the f2fs_record_error_work symbol
in the stack, it can be explicitly a function of f2fs. personal opinion for
reference only.
> Do you mind letting me merge
On 2022/10/25 14:50, Yangtao Li wrote:
Added a new sysfs node called gc_urgent_idle_remaining.
The user can set the trial count limit for GC urgent and
idle mode with this value. If GC thread gets to the limit,
the mode will turn back to GC normal mode finally.
This method is not only
Jaegeuk,
There is a generic fix [1], let's drop this one...
https://lore.kernel.org/lkml/20221025071549.1280528-1-yangyingli...@huawei.com/T/
On 2022/10/20 16:45, Chao Yu wrote:
On 2022/10/18 15:32, Yang Yingliang wrote:
Inject fault while probing module, kset_register() may fail,
if it
https://bugzilla.kernel.org/show_bug.cgi?id=216050
--- Comment #70 from bogdan.nico...@gmail.com ---
Well, if understand correctly, TRIM will still be issued on-demand when
segments are cleared. There's also an option background_gc=sync, which I assume
will block other I/O operations.
https://bugzilla.kernel.org/show_bug.cgi?id=216050
--- Comment #69 from Matteo Croce (rootki...@yahoo.it) ---
That's a terrible workaround.
Does it impact the SSD wear leveling algoritmh?
Can a periodic fstrim mitigate this?
--
You may reply to this email to add a comment.
You are receiving
https://bugzilla.kernel.org/show_bug.cgi?id=216050
--- Comment #68 from bogdan.nico...@gmail.com ---
The only reasonable solution that works for me in everyday use is to deactivate
the background garbage collector entirely. For reference, this can be done
using the background_gc=off mount option
Let's stop threads first to avoid irq context.
Signed-off-by: Jaegeuk Kim
---
fs/f2fs/super.c | 6 +++---
1 file changed, 3 insertions(+), 3 deletions(-)
diff --git a/fs/f2fs/super.c b/fs/f2fs/super.c
index a241e4073f1d..14b4bb0f8ce6 100644
--- a/fs/f2fs/super.c
+++ b/fs/f2fs/super.c
@@
I integrated two patches into single one. I think we don't need Fixes.
On 10/27, Chao Yu wrote:
> On 2022/10/27 18:24, Yangtao Li wrote:
> > Return PTR_ERR(cprc->f2fs_issue_ckpt) instead of -ENOMEM;
> >
>
> Fixes line?
>
> > Signed-off-by: Yangtao Li
>
> Reviewed-by: Chao Yu
>
> Thanks,
On 2022/10/27 20:08, Yangtao Li wrote:
When a fatal error occurs in the file system and it is
remounted to ro mode, the flush thread needs to be stopped.
Fixes: d81ab30e85a5 ("f2fs: support errors=remount-ro|continue|panic
mountoption")
Do you mind letting me merge these two patches into
On 2022/10/27 20:08, Yangtao Li wrote:
Since it is an f2fs module, it is best to prefix function names
with f2fs_. At the same time, there is already a flush thread
and related functions in the system, rename flush_error_work()
to f2fs_record_error_work().
It won't pollute global namespace
On 2022/10/27 18:24, Yangtao Li wrote:
Return PTR_ERR(cprc->f2fs_issue_ckpt) instead of -ENOMEM;
Fixes line?
Signed-off-by: Yangtao Li
Reviewed-by: Chao Yu
Thanks,
___
Linux-f2fs-devel mailing list
Linux-f2fs-devel@lists.sourceforge.net
On 2022/10/27 18:24, Yangtao Li wrote:
Just cleanup for readable, no functional changes.
How about doing such cleanup in one patch?
Thanks,
Signed-off-by: Yangtao Li
---
fs/f2fs/gc.c | 14 ++
fs/f2fs/segment.c | 3 +--
2 files changed, 7 insertions(+), 10
Since it is an f2fs module, it is best to prefix function names
with f2fs_. At the same time, there is already a flush thread
and related functions in the system, rename flush_error_work()
to f2fs_record_error_work().
Signed-off-by: Yangtao Li
---
fs/f2fs/super.c | 4 ++--
1 file changed, 2
When a fatal error occurs in the file system and it is
remounted to ro mode, the flush thread needs to be stopped.
Fixes: d81ab30e85a5 ("f2fs: support errors=remount-ro|continue|panic
mountoption")
Signed-off-by: Yangtao Li
---
fs/f2fs/super.c | 1 +
1 file changed, 1 insertion(+)
diff --git
Need to stop the flush thread running externally,
so export it.
Signed-off-by: Yangtao Li
---
fs/f2fs/f2fs.h| 1 +
fs/f2fs/segment.c | 9 -
2 files changed, 9 insertions(+), 1 deletion(-)
diff --git a/fs/f2fs/f2fs.h b/fs/f2fs/f2fs.h
index b467c0ca1118..0d072b029187 100644
---
Change "truncateion" to "truncation".
Signed-off-by: Keoseong Park
---
fs/f2fs/file.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/fs/f2fs/file.c b/fs/f2fs/file.c
index 8cd52b0965b1..aafdf5fbb825 100644
--- a/fs/f2fs/file.c
+++ b/fs/f2fs/file.c
@@ -571,7 +571,7 @@ void
On Thu, Oct 27, 2022 at 03:47:02PM +0530, Pavan Kondeti wrote:
> On Thu, Oct 27, 2022 at 02:42:41PM +0530, Mukesh Ojha wrote:
> > Data type of msg in f2fs_write_checkpoint trace should
> > be const char * instead of char *.
> >
> > Signed-off-by: Mukesh Ojha
> > ---
> >
Return PTR_ERR(cprc->f2fs_issue_ckpt) instead of -ENOMEM;
Signed-off-by: Yangtao Li
---
fs/f2fs/checkpoint.c | 3 ++-
1 file changed, 2 insertions(+), 1 deletion(-)
diff --git a/fs/f2fs/checkpoint.c b/fs/f2fs/checkpoint.c
index 2a5d9256a6f4..12fd12f2bb97 100644
--- a/fs/f2fs/checkpoint.c
+++
Just cleanup for readable, no functional changes.
Signed-off-by: Yangtao Li
---
fs/f2fs/gc.c | 14 ++
fs/f2fs/segment.c | 3 +--
2 files changed, 7 insertions(+), 10 deletions(-)
diff --git a/fs/f2fs/gc.c b/fs/f2fs/gc.c
index 674a905063db..85d16f4106de 100644
---
On Thu, Oct 27, 2022 at 02:42:41PM +0530, Mukesh Ojha wrote:
> Data type of msg in f2fs_write_checkpoint trace should
> be const char * instead of char *.
>
> Signed-off-by: Mukesh Ojha
> ---
> include/trace/events/f2fs.h | 2 +-
> 1 file changed, 1 insertion(+), 1 deletion(-)
>
> diff --git
Data type of msg in f2fs_write_checkpoint trace should
be const char * instead of char *.
Signed-off-by: Mukesh Ojha
---
include/trace/events/f2fs.h | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/include/trace/events/f2fs.h b/include/trace/events/f2fs.h
index
commit 18ae8d12991b ("f2fs: show more DIO information in tracepoint")
introduces iocb field in 'f2fs_direct_IO_enter' trace event
And it only assigns the pointer and later it accesses its field
in trace print log.
Unable to handle kernel paging request at virtual address ffc04cef3d30
Mem
31 matches
Mail list logo